distilbert base fallacy classification EU AI Act Compliance Profile

distilbert-base-fallacy-classification

This model is a fine-tuned version of distilbert-base-uncased on the Logical Fallacy Dataset.

Model description

The model is fine-tuned for text classification of logical fallacies. There are a total of 14 classes: ad hominem, ad populum, appeal to emotion, circular reasoning, equivocation, fallacy of credibility, fallacy of extension, fallacy of logic, fallacy of relevance, false causality, false dilemma, faulty generalization, intentional, and miscellaneous.

Example Pipeline

from transformers import pipeline

text = "We know that the earth is flat because it looks and feels flat."
model_path = "q3fer/distilbert-base-fallacy-classification"
pipe = pipeline("text-classification", model=model_path, tokenizer=model_path)
pipe(text)

Full Classification Example

import torch
from transformers import AutoTokenizer
from transformers import AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained("q3fer/distilbert-base-fallacy-classification")
tokenizer = AutoTokenizer.from_pretrained("q3fer/distilbert-base-fallacy-classification")

text = "We know that the earth is flat because it looks and feels flat."
inputs = tokenizer(text, return_tensors='pt')

with torch.no_grad():
  logits = model(**inputs)
  scores = logits[0][0]
  scores = torch.nn.Softmax(dim=0)(scores)

  _, ranking = torch.topk(scores, k=scores.shape[0])
  ranking = ranking.tolist()

results = [f"{i+1}) {model.config.id2label[ranking[i]]} {scores[ranking[i]]:.4f}" for i in range(scores.shape[0])]
print('\n'.join(results))

Your obligations as a distilbert base fallacy classification deployer

Track your progress

Sign up to track obligations, receive deadline reminders, and generate audit packages.

Tags

Languages

Library